search

knazeri / edge-connect

EdgeConnect: Structure Guided Image Inpainting using Edge Prediction, ICCV 2019 https://arxiv.org/abs/1901.00212
http://openaccess.thecvf.com/content_ICCVW_2019/html/AIM/Nazeri_EdgeConnect_Structure_Guided_Image_Inpainting_using_Edge_Prediction_ICCVW_2019_paper.html
Other
2.5k stars 530 forks source link

Can not download the pretrained model #109

Open Zealoe opened 5 years ago

Zealoe commented 5 years ago

The link of the pretrained model is unavailable now. Would you please provide a new link? Thanks very much.

AKASH2907 commented 5 years ago

It's working fine and usable too.

zhengbowei commented 4 years ago

hello, AKASH2907. Have you tested the author's CelebA model?

I test the CelebA model using code "python test.py --checkpoints ./checkpoints/celeba --input ./examples/celeba /images --mask ./examples/celeba /masks --output ./checkpoints/results"

and it reports errors loading error: ./examples/celeba/images/celeba_01.png self.to_tensor(img_gray) img_t = F.to_tensor(img).float() img = img.view(pic.size[1], pic.size[0], nchannel) RuntimeError: shape '[256, 256, 1]' is invalid for input of size 262144

I don't know how to solve. thanks you!